Genetic PD Control for Two-Link Manipulator Using Inverse Dynamics

Abstract

Two link manipulators generalize many of the robotic actions and hence have been rigorously studied over the past decade. The path control of the tip of the two link manipulator is a challenging problem due to the coupled, non-linear dynamics. In addition, parameter variations in terms of moments of inertia/mass provide additional challenge to control engineers. To achieve optimal design and operation of the two link manipulator, proper control system design is required. The closed loop control problem of the two-link robot is investigated in this paper, where higher accuracy and high speed of response for positioning the end-effector are required. One traditional control technique, Proportional and Derivative (PD) controller and one artificial intelligence technique, Genetic-PD control (GPD) are applied in this paper. The powerful capabilities of tuning PD controller based on genetic algorithm in locating the optimal solutions to a given optimization problem are exploited by determining the parameters of the PD controller in order to meet specified performance objectives. The performance of the manipulator with the PID is shown to have better dynamic performance as compared to the manually tuned PD. The performance with the GPD is seen to be better in the presence of noise and parameter variations also.